Questions to Determine

• What is the current time offset on the affected node compared to the others?
• Is the NTP/chrony service running on the affected node?
• Can the node reach its configured NTP servers?
• Was the NTP service disabled or misconfigured after the reboot?
• Is there a firewall rule blocking NTP traffic (UDP 123)?
• What does chronyc sources show about NTP server reachability?
• Is the hardware clock (RTC) accurate, or did it drift during the maintenance window?
• What is the maximum clock skew tolerance for etcd and the application?